217 



we adopt the theory that the Valley of the Severn and the 

 escarpment of the Cotteswolds were necessarily formed by the sea, 

 however much the escarpment may resemble a coast-line, we 

 may be idtimately landed in a difficulty which may compel a 

 re-consideration of the subject. 



Whatever may have been the conditions under which the 

 denudation of the Valley of the Severn took place prior to the 

 emergence of the land at the close of the glacial epoch, and 

 the resumption by the Severn of its ancient functions, there 

 remains a large amount of denudation which can only be referred 

 to sub-serial influences, whether by rain, frost, springs, or rivers. 

 To support this opinion it is necessary to refer to the changes 

 which have occurred since the river began to excavate its 

 present channel, and during the period in which marine action 

 has been confined within the same boundary. If we take the 

 height of the cliffs on the banks of the Severn, and assimie, as 

 I think we fairly may, that at some former period the river bed 

 was as high as the top of the cliffs, or nearly so, and that even 

 then the river ran in the lowest part of the valley, we can, by 

 noting the present general level of the surface, and comparing 

 it with the present height of the cliffs, form some idea of the 

 extent of denudation from atmospheric causes. At Fretheme 

 the river has cut a channel some 40 feet deep, reckoning from 

 low water mark, but the general level of the country below 

 Fretherne is nothing Hke 40 feet above the river. The land has 

 therefore been denuded at a rate almost as rapid as the Channel 

 has been excavated, because it being at the present day low- 

 lying laud, it could not have been relatively lower at the time 

 the excavation commenced. The same observations will apply 

 to most of the cliffs forming the banks of the Severn. If the 

 bed of the river were now raised a few feet, its course would be 

 in many places altogether diverted; collateral denudation is, 

 therefore, in my opinion, clearly established. 



In the vaUeys of the Cotteswolds, there is abundant evidence 

 of denudation, but so far as research has hitherto extended, it 

 appears to be sub-serial in its character, facilitated and increased 

 by landslips. I am aware that this opinion does not generally 



